Recent advances in gene therapy of solid cancers have focused attention on the "bystander-effect" in which sensitivity to a therapeutic agent is transferred from genetically sensitive to insensitive cancer cells within a mixed, heterogeneous cancer. Our laboratory has been studying the mechanisms by which clonal populations of tumor cells interact with each other to confer sensitivity to chemotherapeutic agents. Using a mouse mammary cancer model, we have demonstrated significant bystander effects with drugs as pharmacologically diverse as methotrexate, melphalan, thioguanine and cyclophosphamide. We have also demonstrated a diversity of bystander mechanisms, including transfer of drug metabolites via gap junctions, transfer of sensitivity by soluble mediators, cell-cell interactions that alter proliferation rate (hence, sensitivity to chemotherapy) and interactions that utilize host responses. We have prepared genetically engineered mammary tumor cells carrying HSV-TK gene which are sensitive to ganciclovir ((GCV) and have shown that they exhibit a bystander effect in vivo. This proposal is designed to test the hypothesis that (1) similar interactions between genetically engineered and wild type tumor populations are responsible for bystander effects and contribute significantly to chemotherapeutic response and (2) these interactions and their mechanisms are dependent upon tumor stage, site and specific drug regimen. Our aims are:(1) to introduce the GCV-sensitivity conferring, herpes simplex kinase (HSV-TK) gene into others in a series of sister subpopulation mouse mammary tumor cell lines that (a) exhibit bystander effects that work by different mechanisms, (b) differ in immunological and metastatic phenotypes, and (c) differ in intrinsic sensitivity to conventional drugs; (2) to test the mechanisms by which sensitivity to GCV is transmitted to bystander cells, in vivo and in vitro and (3) to develop a strategy for combination chemotherapy of primary tumors and of metastatic disease based on utilization of different bystander mechanisms.